The current model for TJ strands consists of paired rows of Cldn coupled by cis and trans interactions featuring a cis interface derived from crystalline mCldn15 (referred to as X-1) that may not fully account for the natural lateral flexibility and structural variability of TJ strands. X-1 involves a hydrophobic interaction between methionine 68 (M68) and two phenylalanines (F146 and F147). We performed Molecular Dynamics simulations on this strand model and observed an energetic instability resulting in protomer rotations. Also, contrary to what was previously reported, we observed that mutations in M68 do not prevent strand formation. Using live imaging and freeze fracture EM we observed that TJ strands exhibited two distinct lateral flexibility modes: arching, that extends uniformly over a segment of the strand; and bending, that has more abrupt curvature at discrete locations. The estimated angles between Cldn protomers that are needed to account for the arching and bending of the mCldn15 strands range between 0-20o. In addition to strands exhibit other structural variability including various branching modes. The possibilities for structural variability are very broad and suggest multiple forms of cis interactions. To identify new cis interfaces, we used computational protein-protein docking and co-evolutionary coupled mutation analyses. We identified a new cis interface (Cis-1) that shares a common interacting region with X-1 but has a 17o rotation. We observed that independent mutations of several residues involved in Cis-1 or X-1 individually did not prevent strand formation suggesting a level of redundancy between the two interfaces. This functional redundancy in interacting residues between X-1 and Cis-1 may facilitate structural rearrangements where the protomers could rotate or translate without losing contract and cause strand breakage. To validate the Cis-1 interface independently of X-1, we examined mCldn15 R79, a residue unique to the Cis-1 interface. In our Cis-1 model, R79 forms a hydrogen bond with the S67/E157 pair on an adjacent protomer. The requirement of the R79 residue for strand formation and stability was probed using a series of point mutations: R79A, R79W, R79E, and R79H. Mutations of R79 with bulky (R79W, R79H) or negatively charged (R79E) residue completely impeded TJ strand formation, confirming that R79 is critical for TJ strand formation. The general requirement of Cis-1 for TJ strand formation and or stability is supported by our finding that hCldn14 R81, the equivalent to mCldn15 R79, is also essential for TJ strand formation. Interestingly, hCldn14 R81H mutation was reported to be the cause of DFNB29, but the underlying molecular mechanism for this recessive form of hearing loss was not known. Our results indicated that while the mutant hCldn14 R81H can target the plasma membrane, it is unable to form TJ strands. In addition, we observed that hCldn14 R81H does not exert a dominant negative effect on WT hCldn14 as the co-expression of both did not prevent strand formation This observation is consistent with the recessive nature of DFNB29. In the cochlea, Cldn14 strands form the most apical part of the TJ network between sensory hair cells and support cells, acting as the first permeability barrier to the high potassium in the endolymph. The loss of Cldn14 strands likely causes a disruption of the endolymph/perilymph barrier leading to hair cell loss as reported for the Cldn 14 KO mice. We previously reported that non-muscle myosin NMII forms a sarcomere belt that encircles epithelial cells at the level of the apical junctional complex (AJC) and that belts in adjacent cells are connected across the cell boundary forming a transcellular structure. We followed up this work choosing the highly dynamic intestinal epithelium as a model system to contrast with the stable cochlear epithelium, and to further investigate the mechanisms and physiological implications of NMII isoform-specific function. The three mammalian NMII isoforms- NMIIA, NMIIB and NMIIC- each have distinct biochemical and biophysical properties: NMIIA, with low duty ratio and high ATPase activity, is well-suited to driving dynamic rearrangements of actin; conversely, the high duty ratio of NMIIC and NMIIB gear these isoform towards maintaining tension. To localize NMII isoforms at both the tissue and sub-cellular levels of the intestinal epithelium simultaneously, we combined high-performance confocal imaging with multidimensional large image stitching to build several high-resolution maps of cryo-sectioned intestinal epithelial tissue from transgenic mice expressing NMIIA- and NMIIC- GFP. The small intestinal crypt-villus axis represents one of the most dynamic tissues in the body with respect to cell turnover. New cells are generated in the stem cell compartment of the intestinal crypt where they undergo robust amplification and migration along the villus until they are eventually extruded at the villar tips and undergo apoptosis. We found an inverse gradient of NMIIA and NMIIC along the crypt-villus axis, with NMIIC enriched at villi, and NMIIA enriched at the crypts. The increasing levels of NMIIA from crypt to villus correlate with a steady increase in the apical surface area of epithelial cells, a change in cell type from undifferentiated and highly proliferating to differentiated, and change in overall tissue curvature from highly concave to convex. Examining sub-cellular differences in NMII isoform-specific localization we discovered three discrete NMII networks populating the apex of enterocytes: 1) At the AJC, a stable sarcomeric NMIIC network maintains tension across the epithelial sheet, similar to that in the organ of Corti; 2) a novel, transient NMIIA network that appears at the AJC only at sites of cellular rearrangements and extrusion; 3) Both AJC networks connect to a third, two-dimensional, mesh-like NMII network that encompasses the medial apical cortex. This network was further characterized in the stomach where it comprises mini-filaments organized in a striking polygonal array. Using live imaging of acute intestinal and stomach epithelial cultures from the NMIIC-GFP mice, we show that NMIIC within the AJC and medial apical networks is highly dynamic locally, but that the networks remains continuously integrated. Finally, using a conditional knockout mouse for NMIIA, under the villin-Cre promoter to ablate NMIIA expression from the intestine, we show that NMIIC can compensate for the loss of NMIIA sufficiently to maintain organ physiology under normal conditions, despite some defects at the level of cellular morphology. Mutations in NMII isoforms have been linked to a wide number of human pathologies across diverse organ systems, from deafness to macrothrombocytopenia. Collectively, our data provide evidence for differential roles of NMII isoforms in force generation and tensional homeostasis across epithelial sheets. Additionally, we provide new insights into the precise architecture and dynamics of the distinct subcellular actin-myosin networks that coordinate and integrate their functions at the apex of epithelial cells.